INTERNATIONAL SOCIETY FOR

SOIL MECHANICS AND

GEOTECHNICAL ENGINEERING

This paper was downloaded from the Online Library of

the International Society for Soil Mechanics and
Geotechnical Engineering (ISSMGE). The library is
available here:

https://www.issmge.org/publications/online-library

This is an open-access database that archives thousands

of papers published under the Auspices of the ISSMGE and
maintained by the Innovation and Development
Committee of ISSMGE.

The paper was published in the proceedings of the 11th

Australia New Zealand Conference on Geomechanics and
was edited by Prof. Guillermo Narsilio, Prof. Arul
Arulrajah and Prof. Jayantha Kodikara. The conference
was held in Melbourne, Australia, 15-18 July 2012.
 A p-y Approach to Predict the Lateral Load Capacity of Piles
Socketed into Melbourne Mudstone
1 2 2 3
W. L. Chong , A. Haque , P. G. Ranjith and A. Shahinuzzaman
1
Vibropile Pty Ltd, Victoria, Australia; PH (61) 395902600; email: echong@vibropile.com.au
2
Monash University, Department of Civil Engineering, Vic, Australia;
3
Powerlink, Queensland, Australia;

ABSTRACT

Rocks are a highly variable entity where the disparity in composition, depositional history and
weathering condition cause them to have different strength and deformation behaviour. For laterally-
loaded rock-socketed piles, these factors are crucial in determining the lateral capacity of the pile. In
this paper, extensive numerical modelling was undertaken for pile socketed into mudstone rock using
a three-dimensional discrete element package, 3DEC. The numerical model was based on a
calibrated model established by using a full-scale lateral pile load test conducted in mudstone. The
benchmark model was extended to investigate several pre-determined influential rock and pile
parameters and the results were incorporated to form a new p-y criterion for Melbourne mudstone.
The proposed p-y method was subsequently validated using another full-scale pile test in mudstone
and good agreement was observed.

Keywords: Numerical modelling, p-y, mudstone, lateral capacity

1 INTRODUCTION

Many civil engineering infrastructures (e.g., tall buildings, bridges, retaining walls, transmission towers}
are designed to resist high lateral loads generated by wind, waves and earthquakes. The design
methods for laterally loaded piles embedded in soils are well established and calibrated by extensive
full-scale field tests (Meyerhof 1976; Vesic 1977). However, there is a lack of research and field test
data on laterally loaded piles socketed into rocks. Recent advancements have focussed on developing
methods for predicting the lateral load behaviour of rock socketed piles using elastic continuum and
subgrade reaction (p-y) methods (Reese 1997; Yang 2006). A few notable analytical methods (Reese
1997; Yang 2006) have been devised for specific rocks but significant errors were observed when
these methods were applied on different types of rocks (Chong 2011). This is due to the substantial
variation in strength and deformation behaviour of rocks caused by variations in composition,
depositional history as well as weathering condition. To the authors’ knowledge, there has not been
sufficient study conducted or analytical method proposed to predict the lateral capacity of piles
socketed into the local Melbourne mudstone (Chong et al 2008; Chong 2011). Therefore, for more
efficient design solutions, it is imperative to develop an analytical method capable of capturing the
local mudstone’s lateral behaviour.

In this study, three-dimensional distinct element code, 3DEC was employed to investigate numerically
the lateral behaviour of piles socketed into mudstone rock. The numerical model was based on a
calibrated model established for laterally-loaded piles socketed into Melbourne mudstone (Chong et al
2011a; Chong 2011). In this paper, the benchmark model was employed to examine the effect of
several influential rock and pile parameters on the p-y and the corresponding load-deflection
behaviour of piles socketed into mudstone. A new p-y criterion was devised by incorporating the
results obtained from the parametric study. The proposed p-y criterion adopts a hyperbolic form, as
shown in Equation 1, where the curve is governed by Ki (initial subgrade reaction modulus) and pult
(ultimate rock resistance). The p-y criterion was subsequently input into computer program, PYGMY
(Stewart 2000), to estimate the pile head load-deflection behaviour of another full-scale pile test. The
predicted pile head response was observed to match the test results closely.

y
p
1 y
 (1)
Ki pult
Where, p = resultant rock mass resistance per unit pile length (kN/m); y = corresponding ; pile
2
deflection (m); Ki = initial subgrade reaction modulus (kN/m ); pult = ultimate resistance of rock (kN/m)

2012 ANZ Conference Proceedings 596

2 NUMERICAL MODELLING

The numerical model established for laterally-loaded piles socketed into Melbourne mudstone (Chong
et al 2011a; Chong 2011) was employed to carry out parametric study on a host of rock and pile
properties. The effect of these parameters on the initial subgrade reaction modulus (Ki) and the
ultimate rock resistance (pult) of the p-y curve was subsequently analysed to establish a new p-y
criterion for laterally-loaded piles socketed into Melbourne mudstone.

2.1 Initial Subgrade Reactiong Modulus (Ki)

The initial sub-grade reaction modulus for mudstone, Ki, is dependent on a host of rock and pile
parameters. It has been shown by Chong et al (2011b) using laboratory-scale modelling that rock
modulus, pile diameter and pile bending stiffness are the key parameters influencing the load-
deflection response of piles socketed into mudstone. Therefore, this study carried out more detailed
parametric study on these influential parameters. Ki can be mathematically expressed as shown in
Equation 2. The effects of the five parameters were investigated in more detail using the 3DEC model
established based on full-scale pile-load tests.

Ki  f ( Er , D, v, E p I p ) (2)
2
Where, Ki = initial subgrade reaction modulus for homogenous mudstone (kN/m ); Er = rock modulus
(kPa); D = pile diameter (m), v = Poisson’s ratio of rock; EpIp = pile bending stiffness (kNm )
2

Effect of Mudstone Modulus (Er)

Four different mudstone modulus values (62, 220, 540 and 1320MPa) with other parameters held
constant were simulated and the respective p-y behaviour obtained as shown in Figure 1a. The figure
shows the p-y behaviour becomes stiffer with higher rock modulus while the ultimate rock resistance
approaches a similar value. With this, the Ki for each case was then calculated. Figure 1b depicts the
2
variation of Ki with different rock modulus and shows a strong linear relationship (R =0.99). Similar
linear behaviour was also reported by Yang (2006) in his FEM modelling.

Figure 1 a) p-y behaviour and b) Variation of Ki for different mudstone modulus

Effect of Pile Diameter (D)

A number of researchers have found contradictory results for the effect of pile diameter on Ki. Terzaghi
(1955) and Vesic (1961) suggested that pile diameter can be omitted when estimating Ki. However,
Carter (1984) observed from his field test results that pile diameter has a substantial effect on Ki. Yang
(2006) also concluded in his FEM studies on the effect of pile diameter that it has a linear relationship
with Ki. Therefore, an investigation was carried out to study the effect of pile diameter on Ki for
mudstone. The full-scale 3DEC model was extended to simulate three different pile diameters of 0.3,
0.6 and 1.2m while keeping the other parameters fixed. The respective p-y behaviour is plotted in
Figure 3a.

2012 ANZ Conference Proceedings 597

As can be seen from this figure, the p-y response becomes stiffer in both the linear and non-linear
regions with larger pile diameter. The variation of Ki for different pile diameters is illustrated in Figure
2
3b. It can be seen that pile diameter has a strong linear relationship with Ki (R = 0.99). Similar linear
behaviour was also reported by Yang (2006) in his FEM study based on relatively stronger rocks
(shale and sandstone).

Figure 3 a) p-y behaviour and b) Variation of Ki for different pile diameters

Effect of Poisson’s ratio of Mudstone (v)

The effect of Poisson’s ratio on Ki was examined by varying the values from 0.2 to 0.4 based on the
full-scale 3DEC model. Figure 4a depicts the average p-y curves for different Poisson’s ratios
simulated. As can be seen from this figure, the initial linear part of the average p-y curves is identical
for all the Poisson’s ratios investigated. It has also been found by Williams (1980) and Chiu (1981) that
the Poisson’s ratio of the mudstone falls consistently in the range of 0.25 – 0.30 regardless of the
weathering conditions. Therefore, the effect of Poisson’s ratio can be disregarded in the subsequent
derivation of Ki for mudstone.

Figure 4 a) Average p-y curves for different Poisson’s ratios; b) Variation of Ki with different pile
bending stiffness, EpIp

Effect of Pile Bending Stiffness (EpIp)

2
Three values of pile bending stiffness (EpIp) of 3980, 7960, 11940 and 15920 kNm representing pile
modulus of 10, 20 and 40 GPa were simulated while maintaining other parameters constant. The Ki
values for different pile bending stiffnesses were calculated and plotted in Figure 4b. It can be seen
2
from this figure that a power relationship (R =0.99) exists between Ki and EpIp. This finding is
consistent with that of Vesic (1961) as observed for soils and Yang (2006) for relatively hard rocks.

2012 ANZ Conference Proceedings 598

The detailed study conducted on the key influential parameters given in Equation 2 found that
mudstone modulus, pile diameters and pile bending stiffnesses have significant effects on Ki and their
combined effects on Ki can be mathematically expressed by Equations 3-4. A constant (d = 0.3m),
which is the same as the pile diameter tested in this study, is required to normalise the effect of pile
diameter. In Equation 4, the mudstone modulus can be estimated using the well-established water
content relationship (Chiu, 1981) which offers this method a distinct advantage over the existing p-y
criteria (Gabr et al., 2002; Yang, 2006) that require modulus to be calculated using highly subjective
rock classification systems such as RMR and GSI.
0.3
 Ep I p  D  Ep I p 
Ki   ( Er ) ( D)  Ki  Er ( )  
4  (3) d  Er D 4  (4)
 Er D 
Where, A = proportionality constant based on combined effects; d = normalisation constant (0.3m);
2
EpIp = pile bending stiffness (kNm )

Proportionality constant, A, takes into account the combined effect of all the key parameters shown in
Equation 4. The predicted Ki values using Equation 4 are plotted against the Ki derived from 3DEC
models in Figure 5. As can be seen from this figure, the relationship between both initial sub-grade
2
reaction moduli is linear (R = 0.99) with a slope of 1:1. This implies a proportionality value of 1 is
obtained for the given mudstone.

2.2 Ultimate Rock Resistance (pu)

The ultimate rock resistance, pu, for laterally-loaded piles socketed into rock has been investigated by
a number of researchers (Gabr et al., 2002; Yang, 2006). Rock classification systems such as Rock
Mass Rating (RMR) (Bieniawski, 1989) or Geological Strength Index (GSI) (Hoek and Brown 1988)
have been adopted to account for the secondary structures in the rock mass when computing its
ultimate rock resistance. The distinct disadvantage of using these classification systems is the
inconsistency in assigning a rating value for a given rock condition. As it is based on the user’s
experience and judgement, very large discrepancies may occur between different users. Therefore, to
diverge from the unreliable nature of the rock classification systems which have been shown to
produce large under-prediction of the ultimate rock resistance for the local mudstone (Chong 2011), a
different approach is adopted in this study to develop a more practical and reliable method to compute
the ultimate rock resistance, pu.

Figure 5 Comparison of 3DEC and predicted (Equation 4) initial subgrade reaction modulus

This study explored the use of an empirical formula to estimate the ultimate rock resistance, pu, for
laterally-loaded piles socketed into mudstone by taking into consideration the wealth of knowledge
available from previous studies carried out at Monash University on rock-socketed piles in mudstone.
Based on field tests carried out on axially-loaded piles socketed into mudstone, Pells (1980)
suggested the use of a bearing capacity factor with the unconfined compressive strength (UCS) of the
mudstone to estimate the ultimate bearing capacity of the rock. UCS is a well-established property of
mudstone and is widely used in many empirical equations developed for foundation engineering. Pells
(1980) proposed bearing capacity factors of 6 and 12 for mudstone at surface and deeper depth,
respectively.

2012 ANZ Conference Proceedings 599

In the context of laterally-loaded piles socketed into mudstone, the same approach was employed for
predicting the ultimate lateral rock resistance, pu. A series of different bearing capacity factors was
investigated and it was found that a bearing capacity factor of 7 produced the closest and the most
consistent estimation of the ultimate mudstone resistances for different water contents. Therefore,
Equation 5 was adopted for predicting the ultimate rock resistance.

pu  7qu D (5)

The two relationships (Equations 4 and 5) proposed for determining Ki and pult, respectively, were
subsequently incorporated into the general hyperbolic equation (Equation 1) to produce the complete
p-y curve for laterally-loaded piles socketed into mudstone rock. The p-y predictions based on the two
proposed equations are compared with the p-y curves of different mudstone water contents (Chong
2011), as shown in Figure 6a. Figure 6b presents the corresponding pile-head load-deflection
responses for different water contents of the mudstone.

In Figure 6a, it can be seen that the new hyperbolic p-y criterion proposed in this study is able to
capture quite closely the variations encountered in the p-y curves for different water contents of
mudstone. Both the linear and non-linear regions of the computed p-y curves for water contents
ranging from 5% to 20% match very closely the simulated results of 3DEC modelling. Similar good
agreement can be observed in the prediction of corresponding pile-head load-deflection responses for
different mudstone water contents as shown in Figure 6b. In addition to more accurate predictions, the
proposed equation is more practical due to the requirement of only basic rock properties such as
modulus and UCS, which are functions of a single variable, the water content of mudstone, as
opposed to the use of cumbersome correlations with GSI or RMR. The proposed p-y criterion is further
validated against another full-scale pile-load tests conducted in Melbourne mudstone.

Figure 6 a) p-y curves and b) Pile head load-deflection predictions using the proposed hyperbolic p-y
criterion

2.3 Validation of Proposed Method with Full-Scale Pile Load Tests

The results of another full-scale field pile-load tests was utilised to validate the proposed p-y criterion.
The pile test was conducted at the same site as described in Chong et al (2011a) and the water
content of the mudstone was approximated to be 20%. The pile tested had a socket lengths of 1.2m
(L/D = 4). The p-y curves estimated for the given mudstone condition were input into PYGMY (Stewart
2000) to generate the corresponding pile-head load-deflection responses as shown in Figures 8. The
computed load-deflection curves using the proposed p-y criterion show acceptable agreement with the
pile test results. The close predictions reinforce the feasibility and effectiveness of the proposed p-y
criterion to predict the load-deflection behaviour of laterally-loaded piles socketed into mudstone.

3 CONCLUSION

The analysis for laterally-loaded piles socketed into rock is more complex than its soil counterpart due
to the substantial variation in the nature of rocks having different composition, depositional history and
weathering condition which cause the variation in strength and deformation behaviour. Existing

2012 ANZ Conference Proceedings 600

analytical methods for laterally-loaded rock-socketed piles were derived specifically for a given rock
condition and their application on other rock types are questionable and must be treated with caution.
To date, there have not been sufficient studies carried out to develop an analytical tool to determine
the lateral capacity of pile socketed into the local Melbourne mudstone. This study pioneered the
detailed investigation on the lateral behaviour of piles socketed into Melbourne mudstone. A new
subgrade reaction (p-y approach) has been proposed in this paper that can be used in conjunction
with computer package such as PYGMY, to generate the pile head load-deflection response. Through
extensive parametric study, the effects of influential parameters such as rock modulus, UCS, pile
diameter and pile bending stiffness have been incorporated into the proposed p-y criterion. A full-scale
lateral pile load test has been employed to validate the proposed p-y criterion and close agreement
with the predicted pile head load-deflection response has been achieved.

Figure 8 Comparison of pile head load-deflection of predicted pile test

4 ACKNOWLEDGEMENTS

The authors acknowledge the financial support provided by Monash University, Australia and
Powerlink Queensland, Australia for this research.

REFERENCES

Bieniawski, Z.T. (1976) Rock mass classifications in rock engineering. Proceedings of the Symposium on Exploration for Rock
Engineering, Z.T. Bieniawski (editor), Vol. 1, A.A. Balkema, Rotterdam, Holland, pp. 97–106.
Carter, D.P. (1984) A non-linear soil model for predicting lateral pile response. Report No. 359, Civil Engineering Department,
University of Auckland, New Zealand.
Chong, W.L., Haque, A., Ranjith, P.G., and Shahinuzzaman, A. (2008). Lateral Load Capacity of Single Piles Socketed into
Jointed Rock Mass A Review. First Southern Hemisphere International Rock Mechanics Symposium, Vol 1, 297-309
Chong, W.L., Haque, A., Ranjith, P.G., & Shahinuzzaman, A. (2011). Effect of joints on p-y behaviour of laterally loaded piles
socketed into mudstone. International Journal of Rock Mechanics and Mining Sciences, 48(2011), 372-379.
Chong, W.L., Haque, A., Ranjith, P.G., & Shahinuzzaman, A. (2011). A parametric study of lateral load behaviour of single piles
socketed into joint rock mass. Australian Geomechanics Journal, 46(1), 43-50.
Chong, W. L (2011) Lateral Load Capacity of Single Piles Socketed into Melbourne Mudstone. Doctor of Philosophy Thesis of
Monash University, Australia.
Chiu H. K. (1981) Geotechnical properties and numerical analysis for pile design in weak rock. PhD thesis, Department of Civil
Engineering, Monash University.
Gabr, M.A., Cho, K.H., Clark, S.C., Keaney, B.D. and Borden, R.H. (2002) P-y curves for laterally loaded drilled shafts
embedded in weathered rock. Draft Report No. FHWA/NC/2002/08, North Carolina University, Raleigh, NC.
Hoek, E. and Brown, E.T. (1988) The Hoek–Brown criterion – a 1988 update. Proceedings 15th Canadian Rock Mechanics
Symposium, University of Toronto.
Meyerhof, G. G. (1976) Bearing capacity and settlement of pile foundations. Journal of the Geotechnical Engineering Division,
American Society of Civil Engineers, Vol. 102, No. GT3, pp. 197-228
Reese, L.C. (1997) Analysis of laterally loaded piles in weak rock. Journal of Geotechnical and Geo-environmental Engineering,
ASCE, Vol. 123, No. 11, pp. 1010–1017.
Stewart, D. (2000) PYGMY Computer program. Department of Civil and Resource Engineering, University of Western Australia.
Terzaghi, K. (1955) Evaluation of coefficient of subgrade reaction. Geotechnique, 5(4), pp. 297 – 326.
Vesic, A. S. (1977) Design of Pile Foundations. National Cooperative Highway Research Program Synthesis of Practice No. 42,
Transportation Research Board, Washington D. C.
Vesic, A. S. (1961) Beam on elastic subgrade and Winkler hypothesis. Proceedings 5th International Conference Soil
Mechanics and Foundation Engineering, Paris, Vol. 1, pp. 845–850.
Yang, K (2006) Analysis of Laterally Loaded Drilled Shafts in Rock. Doctor of Philosophy Thesis of University of Akron, United
States.
Williams AJ. (1980) The Design and Performance of Piles Socketed into Weak Rock. Doctor of Philosophy Thesis of Monash
University, Australia.

2012 ANZ Conference Proceedings 601